This study demonstrated that graduate applicants compared to school leaver applicants were significantly more conscientious, more confident, more self controlled, more communitarian in moral orientation and less anxious. Only one of these differences was preserved in the entrants with graduates being less anxious. However, the graduate entrants were significantly less empathetic and conscientious than the school leavers.
The principles of 'widening access' to higher education are accepted to be sound in principle but evidence of the value of the approach is lacking both in general [19
], and specifically in respect of selecting health professionals [8
] and medical students [20
]. Our study was an attempt to examine whether widening access by the strategy of Graduate Entry to medicine, increased psychological diversity and that such an increase, if found, was desirable. In principle, the goal is to recruit students who can cope successfully with the demands of the course and become caring and competent doctors. In particular they should be less rather than more susceptible to depression, anxiety and stress. We would like entrants to have psychological profiles that are not at the extremes of the scale. Diversity in personality in the medical workforce is important as there is evidence that different specialty choices may be influenced by personality traits [21
We acknowledge that the study has limitations, mainly related to study numbers. The sample sizes in some subgroups limited the comparisons that could be made. For example, only seven applicants to the 4 year graduate entry course and 30 to the 5 year school leaver course completed the tests at both application and entry. The ideal would have been to have sufficient numbers to allow a full 'within subjects' comparison (tested at application and entry) by course type (5 or 4 year course) and to also consider those who completed the tests at application but were not offered a place. The logistical problems of increasing the numbers especially in the applicants should not be underestimated. Increasing the number of participating medical schools may be one approach but it would introduce another confounder to the analysis and the number of students from an individual medical school may not be increased.
The key question for the interpretation of the results is how representative are the groups participating in the studies. We addressed this by undertaking subgroup analysis using the same academic and socio-demographic criteria we reported previously [4
]. The school leaver applicants who participated were significantly younger (though the median age was 18 y in both groups), had a higher proportion of white individuals (67% vs 60%) and higher prior academic achievement (Median UCAS Tariff Points 450 vs 360; UCAS Tariff Points are awarded to students on the basis of their achieved A-level grades - ie A = 120 points, B = 100 points etc) than those who did not participate. Thus, the school leavers who participated were different by academic and socio-demographic criteria and so it might be speculated that they would respond differently to the psychological testing. In contrast, there were no significant differences in the academic and socio-demographic criteria between the graduate applicants who participated and those that did not and thus the argument of them being unrepresentative is less strong. It is even more difficult to make the case for the school entrants who participated being unrepresentative as 98% of the year group participated. However, the potential for bias still exists in the 39 4-year entrants as they only comprised 43% of the total. Whilst this group had the same academic and socio-demographic profile typical of all graduate entrants which we have reported previously [4
] and in that respect could be considered 'representative' there may be other unidentified ways in which such a small sample could be biased and 'unrepresentative'.
In our view, the reasons for the low participation rates in the applicants is most likely to be due to the requirement of applicants to come to one of four centres in the Christmas holidays to take the test. It had no bearing on their application and those who participated did so out of goodwill. The high compliance rate amongst the school leaver entrants we believe was explained by the timetabling the tests in the first month of their course. Even though the voluntary nature of the study was stressed arguably it was too early for them to have realised they could exercise that choice. In contrast, we speculate that the experienced graduates, who also had a timetabled session for the tests, strategically chose not to take the test as they were more focussed on their studies rather than research.
Despite these concerns over participant numbers, we are of the view that the sample sizes in this study did allow a reasonably meaningful comparison between school leaver and graduate applicants. The relevance of the comparisons between school leaver and graduate entrants is lessened by the small numbers in the latter group. However, it should be noted that studies that have compared differences in personality scores in selection contexts have generally found that the same pattern of results is observed for between subjects designs (as here) and within subjects designs [23
]. In fact the observed effect sizes are generally smaller for between subjects designs [23
]. Similarly the test/retest coefficients generally indicated acceptable temporal consistency with level of stability at or greater than averages for similar traits reported in recent meta-analytic studies [24
]. Overall, therefore, we feel can have some confidence in the validity of our findings for the differences between the two groups of applicants.
The differences between the two groups of applicants were perhaps not surprising with the graduates being more conscientious, more communitarian, more confident and less anxious - all attributes one might expect from individuals who have successfully complete a higher degree. In addition, they are consistent with the reported changes in personality with age and the fact that the graduates are older [25
]. However, it should be noted that, whilst statistically significantly different, the actual differences are not great and likely to be somewhat subtle in terms of actual observable behavioural differences.
Of these psychological traits, the only one that was also more common in the graduate entrants was that they were less anxious. However, given the small numbers in the graduate entry group it is difficult to interpret the relevance of the differences between them and the school leaver entrants. Specifically the 39 graduates who participated were less agreeable, less conscientious, more aloof and less empathetic. Given that the 40 minute structured interview that the graduates undertake is, in part, designed to select individuals who are agreeable, conscientious, less aloof and more empathetic we are concerned by the potential significance of this finding. However, whilst the differences may be small and subtle in terms of meaningful behavioural differences, we need to undertake the study with larger and more representative graduate entrants to be sure of these findings.
The importance of measuring personal qualities in medical students has been debated frequently. We have suggested that the PQA might be a useful tool in medical student selection, though, to our knowledge, there have been no reports of actual use of psychological personality testing in selection in practice [1
]. In contrast, there have been a number of reports that have shown a correlation with certain personality types and progress on the medical course. Manuel et al
] using the 16 Personality Factor Questionnaire showed a significant correlation to clinical skills performance in 206 second year medical students. Specifically, the clinical skills score positively correlated to 'warmth' and negatively with 'abstractedness' and 'privateness'. Communication skills correlated positively with 'warmth', 'emotional stability' and 'perfectionism' and negatively with 'privateness' [19
]. We surmise that 'warmth' may be somewhat equivalent to Empathy on the NACE test used here, and 'privateness' to Aloofness. We have shown that personality, and the domain of 'conscientiousness' in particular, was positively correlated to performance across the whole 5-year course in Nottingham medical students [8
]. However, we also showed that this effect was strongest in the preclinical years and became less in the clinical years, and in a structural equation model there was a negative correlation between conscientiousness and clinical skill performance. The positive effect of conscientiousness has been shown by others too. In a study of 785 medical students, Lievens et al
. demonstrated that conscientiousness significantly predicted end of year results in the preclinical years [20
]. Furthermore, medical students who scored low on conscientiousness and high on gregariousness and excitement-seeking were shown to be significantly less likely to be successful in their exams [20
We are unaware of any studies of personality and performance of graduates specifically. However, there have been several which demonstrate that personality influences the choice of career specialty in graduates. Petrides & McManus (2004) followed three large cohorts of students into their postgraduate careers and studied their choice using Holland's RIASEC (Realistic-Investigative-Artistic-Social-Enterprising-Conventional) Typology of careers [21
]. Typical associations found were Surgery (realistic), Hospital Medicine (investigative), Psychiatry (Artistic) Public Health (Social), Administrative Medicine (Enterprising) and Laboratory Medicine (Conventional) [21
]. Stilwell et al
. studied nearly 4000 US medical graduates using the Myers-Briggs Type Indicator. They found that 'Feeling types' (in contrast to 'Thinking types') chose family medicine, while 'Thinking types' chose surgical specialties. Women were more likely to be 'Feeling types' and men 'Thinking types' [22
These studies of personality and career choice suggest that we should be admitting students and graduating doctors suited to a range of careers with a variety of personality types though not extreme. In our earlier studies, we showed that graduate applicants brought greater diversity in socio-demographic (being older, more socio-economically deprived and more males), academic (lower prior academic achievement) and non-academic areas (a variety of personality differences and fewer themes in UCAS personal statements and references) [4
]. In the selection process some of this diversity was retained (age, gender, socio-economic, academic and UCAS statements) but some was lost. However, that study acknowledged that the diversity which graduates brought did not result in a student population that was a mirror image of society and certainly that is not the goal of 'widening access'. In the current study also we have shown that graduate applicants brought a greater psychological diversity. The paper reports that the main elements of this diversity that were lost in the selection process were the differences in the personality traits and as a consequence a narrower spectrum of applicants was selected. If this finding is a true representation of the whole cohort, it is interesting to speculate whether this is an advantage or a disadvantage.
We are currently studying the impact of the GEM students on the course. The GEM students and the 'school leaver' group have different 'preclinical' training, but they are merged and mixed in their clinical years. The present study has shown that some of these desirable characteristics are more prevalent in the 4-year entrants than in the 5-year students.